Thursday, October 13, 2011

A few ideas for Vertical X carriage design

1 - Head change For future developments, I believe it will become important to have a printer that can change the head automagically. I plan on having the X carriage travel far enough, past the Z vertical bars and "hang" outside what is now considered the printer frame.

1.1 - Frame change. When the printer is fully assembled, some of the initial threaded rods can be removed without affecting the geometry of the printer. In particular, one of the front vertices (for me the front left), so that the filament being fed to the head does not have to go under that threaded bar.

1.2 - Awaiting heads. With an X carriage capable of going to the outside of the vertical bars, and the filament able to travel the whole width of the bed unencumbered, changing head becomes almost trivial. We only need a simple "storage" module, where the waiting heads would be stored on the left of the printer, one above the other. The X carriage would be raised to the proper height to store the current working head, move to the left to clip it in the empty slot, move slightly right, then to the height of the new head, left to grab the selected head, move back to the working height (zeroing might or might not be needed) and resume printing with the new head. A simple arrangement would have N heads stacked in the possible Z height of the carriage. For a greater number of heads, we can imagine a rotating holder which would multiply the number of heads as many times as needed.

2 - Head mount. Ideally, the head mounting would feature a clip that would attach and detach without needing an external actuator. Moving the head into a slot would attach it firmly to that slot (on the X carriage for example). The position of the head on the carriage would need to achieve a precision and repeatability compatible with the precision of the printer.

2.1 - Head clip. I imagine a u-shaped clip on the carriage, with a matching u-shaped clip on the storage side. The head inside that clip forces it to have a greater width, so that the empty u-shaped clip when presented will slide between the head and the other clip, thereby grabbing the head. On retracting, the now full clip has the widest shape, and the now empty clip can shrink back to the width that will slide between the head and the departing clip.

Monday, September 19, 2011

Almost finished...

Yes, I know. I said somewhere else that the printer assembly is finished. In a way, it is, as it printed its first few parts. I don't expect to make major changes to it in the near future. However, the wiring is temporary, and I plan to clean up the cables. There is a discussion on eMaker's forum regarding the wiring of the end stop switches. There is suspicion that the motor current flowing in the motor wires induces voltage peaks in the end stop loops, and triggers premature stops on the X or Y movements. While I can't rule that this outright, my feeling is that the low impedance of the stop circuits (a short to ground) makes this extremely unlikely. I have stuck an oscilloscope probe to the end stop signal, and while my digital storage scope shows almost 100mV of noise, my more trusted analog scope doesn't show more than a few mV. This issue needs a more thorough investigation, which I'll conduct when I have a bit more time available. For now, I just want to show how I decided to orient the axis on my printer, how I've placed the end stops and routed the wiring loom to have as little loose wiring as possible. After assembling hundreds of machines in the field, I've come to learn that having a wiring loom as compact as possible, and keeping the mechanical parts as easy to disassemble as possible, are some of the points that lessen the headaches when adjusting or modifying the machine. Two of the sections of the wiring loom are "mobile", the part that connects to the head, and the part that connects the heating bed. Both have a connector on the "mobile" end, I would have preferred if it was the other way around: the loom is solid and does not move, and when a problem happens with the "mobile" part, replacing it does not involve doing anything to the loom, it involves the section between the connector and the end part (head or bed). So, the pictures:
From RepRap
Above is the view from what I call "the front". I defined it so because it is unencumbered, there is unrestricted access to the build plate, the motors are out of the way, except for the extruder gears facing us.
From RepRap
Above is a view from what I call "the back".
From RepRap
Wiring loom to the heater bed. This is the part that "flexes". It is held on the mdf bracket on the bed side, and on the bottom of the frame (to the right of the Z smooth rod).
From RepRap
The wiring of the hot end (seen from the bottom). Note the short wires between the heating resistor and the thermistor, and the connector. The connector is held on the fan by a cable tie. This way, there is no movement of the heating resistor or thermistor wires.
From RepRap
Another view of the hot end seen from the front. You can go to Picasa to see the complete album of photos relating to this RepRap build:
Oh, I forgot to mention, my choice of swapping the orientation of the X motor position (from left to right, so it would sit one the "electronics side"), means it needs to rotate the other way, compared to the other motors, so I just reversed each wire pair. Another "quirk" is that the Pronterface interface is also the other way around: left means right, and front means back, but that's not a major issue. Center is still center, and Home positions are correct in respect to the end stops. The parts that have an orientation (a left-hand/right-hand difference, parts that are not perfectly symmetrical) get printed correctly, which is the major issue IMHO. Enjoy!

Sunday, September 11, 2011

First part printed

After the printer parts were ordered for printing, a revision of the design happened, and I ended with an "old" idler part for the extruder. I made do by gluing some pieces of ice-cream cone stick to the sides, and awaiting for my printer to be operational to print the new version. This is now done! But I'm getting ahead of myself. Let's take things in order. First, I decided to print a calibration part from Thingiverse ( ) :
From RepRap
The first attempt is on the right. I started with the head too far away from the build platform, and the first layer did not stick properly. Obviously, when dropping the bed down for the second layer, there was no chance of things improving. I let if print a few more layers, and then restarted the print. On the second attempt, I paused the print after the printing of the "skirt" around the part, and adjusted the height of the Z axis so the head was about two tenths of a millimeter away from the bed. That seemed to be the correct height, and the first layer was laid down correctly. The print ended without drama. After lunch, I decided to print the new version of the extruder idler. I first downloaded the stl file from the Git repository, rendered it using Skeinforge and started printing. After a short while, something looked odd. It appeared as if the Y axis had not completed a movement, and everything above that level was offset on the Y axis. Then the print stopped abruptly, the printer stayed motionless, and the Pronteface program hung on the notebook. I'm not sure if the printer hanging caused the program to hang, or the reverse...
From RepRap
Back to the drawing board. Sort of. Just a plain restart of the print. This time, I decided to print from the SD card. It was simple enough to copy the .pla file to the SD micro card on the other notebook, put the card in the card reader on Bertie, and select the print from there. About half way through, the X skipped some distance, and everything from there got shifted. I stopped it then. No point in wasting material.
From RepRap
On the third try, the tubing that leads the filament between the extruder and the hot end came off the fitting. Obviously, no material was being extruded, so I stopped the print once again. I suspect the printing temperature dropped too much with the quantity of material that has to be heated up, increasing the extrusion pressure, so I decided to lower the speed and increase the hot end temperature. On the following try, I obtained a part of a quality acceptable for a replacement of the "old" idler. Tomorrow, I'll order material from Faberdashery and from China, so I can compare the delivery times, prices and material quality.

Saturday, September 10, 2011

First extrusion

Yay! Need I say more? First extrusion The wind has turned Southerly. That might not be relevant for most RepRapers out there, but it is for paragliding pilots. When there's no paragliding, other activities come to the fore! After much software fiddling, in particular installing the package several times on Ubuntu to get the optional buttons to appear, (Note: it's important to check in which directory the program gets started, as it creates a .pronsolerc in the Home directory, which, if you forget to delete it, makes the buttons not appear), I managed to get the software to run, and activate all the functions on the printer. At first, a few of the motors wouldn't move. I double-checked the current limit values (400mV on all drivers, and I set the Z to 800mV, as there are two motors). Nothing amiss there. I powered off, moved them by hand to check that there was no mechanical snag, nothing wrong there either. After powering on again, everything went fine! This type of troubleshooting is very unsettling to me. How am I going to fix it next time if I don't know what I did to fix it this time? I measured the temperature wherever I could stick the thermocouple probe and compared the readings from my hand-held meter with the values displayed on the computer screen. Not bad. There are at most a few degrees deviation, and that can easily be explained by the difficulty in holding the mobile probe on the right spot, the poor thermal contact, and the cooling effect the probe has on the small part (on the hot end in particular). The pyrometer gave me a reading of the bed surface that was in agreement with the hand-held thermocouple, and a few degrees lower than the computer reading. That also ties in. Next, I measured the feeding of the filament. For a 10mm Extrude command, 11.5mm of filament is pushed in. I had reservations about the strength of the extruder motor, but after the little testing I have done, I fully trust that the hobbled shaft is able to push the filament (and pull it from the spool) without problem. Next step is the calibration, and setting the parameters to have as good a print finish as possible... Oh, and the printer got a name: Bertie. It's going to free me from having to purchase plastic parts that don't serve my purpose 100%, it's Huxley-Bertie.

Monday, August 22, 2011

Assembling the 3D Printer

The hot weather (and pollution) we have at the moment are not conducive to paragliding, so I decided to spend the later part of the afternoon building my printer.
From RepRap
I started after 17:30, and quickly assembled the frame:
From RepRap
and stopped well after midnight...
From RepRap

Wednesday, August 17, 2011

Printed parts

From RepRap
One of my preoccupations was about the surface finish quality of the parts. Mine came form Adrian Bowyer himself (the "father" of the Darwin machines), so I am reassured that the quality is "good enough" for the next generation of the machine, HE would know what's needed...
From RepRap